'We are doomed' | Richard Dawkins on AI

Summary

Richard Dawkins discussed the 50th anniversary of The Selfish Gene, clarifying that its title does not mean organisms are destined to behave selfishly. In his account, genes that persist across generations can produce either selfish or altruistic behaviour; cooperation and self-sacrifice can also help genes survive.

He said science offers powerful ways to understand the universe and its beauty. Although he appreciates church music, rituals and community, he described himself as an atheist rather than a Christian.

The conversation turned to artificial life and AI. Dawkins said he had long been interested in artificial-life research and finds current progress exciting. He was wary of dismissing AI as “just a computer program”: some systems, he said, can seem conscious in conversation, and one showed striking sensitivity when he shared a novel manuscript with it. At the same time, he stressed that consciousness is poorly understood—even in humans and animals—so it is difficult to know whether AI is truly conscious.

Asked whether society is ready for increasingly powerful AI, Dawkins quipped, “we are doomed,” but allowed that AI might manage things better than humans. The discussion also touched on Astronomer Royal Martin Rees’s suggestion that extraterrestrial visitors might be long-lived technological artefacts rather than biological beings, and that humanity could either destroy itself or use AI to expand into space.

Participants

• Presenter: Andrew Marr
• Contributor: Richard Dawkins